Species-Specific Genes Under Selection Characterize the Co-Evolution of Slavemaker and Host Lifestyles B

Feldmeyer et al. BMC Evolutionary Biology (2017) 17:237 DOI 10.1186/s12862-017-1078-9

RESEARCHARTICLE Open Access Species-specific genes under selection characterize the co-evolution of slavemaker and host lifestyles B. Feldmeyer1* , D. Elsner2, A. Alleman3 and S. Foitzik3

Abstract Background: The transition to a parasitic lifestyle entails comprehensive changes to the selective regime. In parasites, genes encoding for traits that facilitate host detection, exploitation and transmission should be under selection. Slavemaking ants are social parasites that exploit the altruistic behaviour of their hosts by stealing heterospecific host brood during raids, which afterwards serve as slaves in slavemaker nests. Here we search for evidence of selection in the transcriptomes of three slavemaker species and three closely related hosts. We expected selection on genes underlying recognition and raiding or defense behaviour. Analyses of selective forces in species with a slavemaker or host lifestyle allowed investigation into whether or not repeated instances of slavemaker evolution share the same genetic basis. To investigate the genetic basis of host-slavemaker co-evolution, we created orthologous clusters from transcriptome sequences of six Temnothorax ant species - three slavemakers and three hosts - to identify genes with signatures of selection. We further tested for functional enrichment in selected genes from slavemakers and hosts respectively and investigated which pathways the according genes belong to. Results: Our phylogenetic analysis, based on more than 5000 ortholog sequences, revealed sister species status for two slavemakers as well as two hosts, contradicting a previous phylogeny based on mtDNA. We identified 309 genes with signs of positive selection on branches leading to slavemakers and 161 leading to hosts. Among these were genes potentially involved in cuticular hydrocarbon synthesis, thus species recognition, and circadian clock functionality possibly explaining the different activity patterns of slavemakers and hosts. There was little overlap of genes with signatures of positive selection among species, which are involved in numerous different functions and different pathways. Conclusions: We identified different genes, functions and pathways under positive selection in each species. These results point to species-specific adaptations rather than convergent trajectories during the evolution of the slavemaker and host lifestyles suggesting that the evolution of parasitism, even in closely related species, may be achieved in diverse ways. Keywords: Positive selection, Social parasites, Temnothorax,Co-evolution

* Correspondence: [email protected] 1Senckenberg Biodiversity and Climate Research Centre (BiK-F), Molecular Ecology, Senckenberganlage 25, 60325 Frankfurt am Main, Germany Full list of author information is available at the end of the article

© The Author(s). 2017 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Feldmeyer et al. BMC Evolutionary Biology (2017) 17:237 Page 2 of 11

Background invade other colonies and reproduce clonally [36]. Iden- Parasitism is one of the most successful modes of life, as tified candidate loci for this form of social parasitism in- measured by how often it evolved and by how many clude genes involved in ecdysteroid signalling, juvenile parasitic species presently exist [1]. Parasites are a taxo- hormone and dopamine biosynthesis, which may regu- nomically highly diverse group, and range from intrage- late worker ovary activation [37]. A study on three work- nomic ‘genetic’ parasites, through microparasites erless inquiline social parasites of Vollenhovia ants in (viruses, bacteria and protozoa) and macroparasites comparison to their Pogonomyrmex hosts found little (worms, arthropods and even vertebrates) [2], to brood evidence for gene loss, damaging mutations, or shifts in and social parasites [3, 4]. selection regimes, suggesting that regulatory changes, ra- Parasites are ideal biological models for the study of ther than sequence differences play a role in the evolu- ecological specialization, speciation mechanisms, diversifi- tion of these workerless social parasites [35]. However, cation and co-adaptation [5]. The relationship of hosts the genomic basis of the slavemaker lifestyle and its’ pe- and parasites is one of mutual adaptation with parasites culiarities has never been investigated. trying to dupe the host, whereas hosts adapt to defend Here we explore the evolution of the slavemaker life- themselves [6–12]. Co-evolutionary dynamics are shaped style in North American Temnothorax ants, a taxon in by numerous factors including life history traits [13, 14], which slavery evolved several times independently [38]. epidemiological characteristics [15, 16], population size We specifically focus on three slavemaker species T. [17, 18], fluctuating environmental changes [19], the pres- americanus, T. duloticus and T. pilagens, and their three ence of multiple parasites reviewed by [20], and social in- closely related host species, T. longispinosus, T. curvispi- teractions within the host taxon reviewed by [21, 22]. nosus and T. ambiguus [39, 40]. The dulotic lifestyle of Signatures of balancing selection are expected on im- these three slavemakers is characterized by recurrent munity genes, playing a major role in the co-evolution and destructive slave raids during summer [39]. During between micro-parasites and their hosts, whereas genes these raids, slavemaker worker raiding parties search for encoding behavioural or morphological traits, important and attack host colonies to steal worker brood. Upon in social parasites and their hosts, should show signs of their emergence as adult workers in slavemaker nests, positive selection [23]. the social behaviours of these enslaved host workers will Social parasitism, is a special form of parasitism, where be exploited by the slavemakers, whose workers lost the the social behaviour of the host, rather than its physi- ability to care for themselves [41]. While host nests on ology, is exploited [3]. Avian brood parasites, such as average contain around 50 workers [42], the number of cuckoos and cowbirds take advantage of the brood care workers in slavemaker nests is much lower with on aver- behaviour of other bird species, and thus avoid the costs age approximately five workers [39, 40, 43]. Moreover, of parental care [24, 25]. Several avian brood parasites slavemaker workers are only active during the raiding evolved from non-parasitic ancestors, and started out by season and do not take over normal worker chores such exploiting the brood care behaviour of their conspecifics as brood care and foraging [44, 45]. Each slavemaker [26]. Similarly, ant social parasites, can arise via sympat- species exhibits distinct morphological characteristics ric speciation from their later host [27]. Such a transi- (e.g. size and colour), and raiding behaviours [41, 44]. T. tion to a parasitic lifestyle should lead to the selection of americanus - the most derived parasite in the group in traits important for a parasitic mode of life, such as host terms of morphology and behaviour - mainly uses a recognition, circumventing the host defence system, and propaganda pheromone to induce panic among hosts, transmission. Indeed, avian brood parasites lost their preventing organized evacuation or nest defence [41, 46, ability to build nests [28], a social parasitic wasp needs a 47]. The strategy of T. pilagens is quite variable, and specific host species to be successful as parasite [29], may also depend on the aggressiveness of the host col- and many slavemaking ants are unable to even feed ony [40, 48]. In some instances host workers are killed themselves and completely rely on the care of their by stinging, while in other cases the raid is seemingly enslaved host workers [30]. peaceful without any casualties, facilitating the incorpor- We are just starting to understand the genomic basis ation of even adult host workers into the slavemaker col- of the parasite lifestyle as such [31–33 and authors ony [48]. T. duloticus is a fierce slavemaker that mostly therein], and some first patterns of convergence, gene stings all opponents to death before taking the brood, losses and gains become apparent [33 and authors resulting in the local eradication of host colonies [41, 43, therein, 34]. First studies on the genetic basis of social 44, 49]. Each of the three slavemakers can exploit several parasitism concentrated either on ant inquilines - social host species, but has a clearly preferred host. The de- parasites, that secondarily lost the worker caste and in- rived T. americanus uses all three Temnothorax species, quiline queens live within host colonies [35] or the Cape but focusses when possible on T. longispinosus [50]. T. honeybee (Apis mellifera capensis), in which workers duloticus occasionally attacks T. longispinosus but Feldmeyer et al. BMC Evolutionary Biology (2017) 17:237 Page 3 of 11

prefers T. curvispinosus [43] and T. pilagens prefers T. genes in respect to raiding and nest defence behaviour, ambiguus over T. longispinosus [40, 48]. we obtained transcriptomes of ants engaged in these re- Co-evolution between the obligate social parasites and spective behaviours. Six workers per species and behav- their hosts not only leads to adaptations in slavemakers, iour were pooled for RNA isolation in replicates of six. but also to counter-adaptations in behavioural, chemical Libraries were constructed and sequenced paired-end on and life history traits in host species and populations an Illumina HiSeq 2000 at GENterprise Genomics. Se- [11, 47, 51–54]. Host aggression [54], as well as host de- quences were quality trimmed with Trimmomatic v0.32 fence strategies [55] are linked to geographic variation in [62]. De novo assembly of the transcriptomes was con- parasite pressure. It is known that adaptations to similar ducted using a combination of the CLC bio workbench ecological conditions may lead to the evolution of simi- (Qiagen) and MIRA [63] (Additional file 2; for more de- lar (convergent) phenotypes in non-related species. The tails see [64]). Contigs were annotated using BlastX degree to which parallelism extents to the molecular v.2.2.30 against the non-redundant arthropod database level has recently experienced an upsurge of interest (November 2014). The online tool ORFpredictor 2.0.3 [56–60]. Evidence is ambiguous, with some studies [65] was used to predict open reading frames and amino pointing to parallelism, and others to species-specific acid sequences for all contigs. The predicted and trans- trajectories [59, 61 and authors therein]. Moreover, it be- lated amino acid sequences were used as input for comes clear that the level of organisation plays a major OrthoMCL 2.0.9 [66], to build ortholog sequence clus- role in detecting convergent evolution, as the degree of ters. In total we obtained 55,521 orthologous protein parallelism is predicted to increase from the nucleotide clusters, out of which 6432 clusters contained at least level to features of whole organisms [61]. The North one sequence per species. These clusters were filtered American Temnothorax system, with six closely related with an in-house python script (available from GitHub: slavemaker and host species is ideal to study the genetic https://zenodo.org/record/60135#. V9k495h96Uk) based basis of repeated evolution of phenotypic traits involved on pairwise Blast similarity scores, which resulted in in host-parasite co-evolution. 5791 clusters with a single sequence per species. After The main objective of this study was to investigate the trimming these sequence alignments with Gblocks 0.91b selective forces shaping the host and slavemaker life- [67], 5199 clusters remained for further analyses (NCBI styles, and the organisational level of convergence. The Bio Project GSE95604). main questions we tried to answer were: Which genes are under positive selection in slavemakers or hosts? Is Phylogenetic analysis molecular parallelism involved in the convergent evolu- We chose the myrmicine ant Acromyrmex echinator to tion of slavemaker lifestyles? Do we find convergence on include as outgroup, for which we observed the highest the gene, functional or pathway level? Blast similarity in our contig Blast searches. A. echinator sequences were obtained from the “Hymenoptera Gen- Methods ome Database” [68; aech_OGSv3.8_pep.fa]. We inferred Ant colonies were collected over the course of 2 years orthology between T. curvispinosus and Acromyrmex (2013 and 2014) in New York State, Ohio and Michigan echinator applying a local BLASTn [69]. The according (Additional file 1), and brought back to the lab in Mainz. sequences for each cluster were obtained and aligned To induce raiding activity, colonies were moved during with Mafft 7.0 [70]. The alignments were trimmed using the raiding season in August to 25 °C, 14 L:10D light Gblocks 0.91b [66] with default settings. All clusters cycle conditions 1 week prior to the onset of the raiding were concatenated into a single alignment and the pro- experiment. Raiding arenas (30 × 40 cm plastic boxes gram ProtTest 3.4 [71] was used to calculate the with plastered floor) were set up, in which each slave- appropriate evolutionary model (JTT + I + G + F). A maker species was allowed to raid colonies of its pre- Maximum Likelihood phylogenetic tree with 1000 ferred host species. We waited until slavemaker scouts bootstrap replicates was constructed with RAxML 8.1.16 had returned to their mother nest and recruited add- [72]. We additionally estimated evolutionary models for itional raiders to infiltrate the host nest, and aggressive each single cluster and constructed the respective encounters between slavemaker and host workers could Maximum Likelihood trees for the codeml analyses (see be observed. This was the time we sampled workers ac- below). tively engaging in a raid or nest defence respectively. To obtain workers in a somewhat neutral behavioural state, Tests for positive and relaxed selection we collected host individuals outside the nest before To test for signatures of positive selection the software raids, as well as slavemaker workers outside the raiding package PAML 4.8 [73] was used to apply the branch- season under the same external conditions. Since we site model A in codeml (model = 2, NSsites = 2). codeml were interested in the evolution of slavemaker and host estimates the nonsynonymous/synonymous substitution Feldmeyer et al. BMC Evolutionary Biology (2017) 17:237 Page 4 of 11

ratio (ɷ = dN/dS), where ɷ = 1 indicates neutral evolu- Results tion, ɷ < 1 purifying selection, and ɷ > 1 indicates Phylogenetic analysis positive selection. To test for statistical significance log- Based on the concatenated sequence of 5199 ortholog likelihood ratios were calculated and FDR corrected for sequence clusters the RAxML phylogenetic analysis re- multiple testing [74]. The cluster specific tree topology, sulted in a tree with well supported nodes (Fig. 1). T. as inferred by RAxML was used as input for codeml.To americanus is the most distant taxa to the other five test for positively selected genes, we coded each single Temnothorax species, as corroborated by its deviant species as foreground branch, and additionally the set of phenotype and a previous phylogenetic analysis [38]. In slavemaker branches as well as the host branches contrast to this earlier phylogeny based on two mito- respectively. chondrial loci [38], our nuclear tree now supports a sis- The online tool Venny 2.1 (http://bioinfogp.cnb.c- ter species relationship between the two younger sic.es/tools/venny/) was used to visualize shared and slavemaking species T. duloticus and T. pilagens [40], as species-specific genes. To statistically assess to what ex- well as between the two host species T. ambiguus and T. tent the observed intersection in divergent features longispinosus, with T. curvispinosus being the next dis- among pairs of species would be expected by chance, we tant taxon, followed by T. americanus. applied a randomisation procedure implemented in a custom Python script. We used 10,000 replicates to infer Positively selected genes how often an observed intersection of size i of x and y In total, we found 574 genes under positive selection; positive draws from a base population of size z was lar- 309 on the branches leading to slavemakers, and signifi- ger or smaller than those from random draws. cantly less, 161, positively selected genes on the 2 branches to hosts (χ1 = 77.85, p < 0.0001; Add- itional file 3). Looking at the branches of each single Enrichment analyses species, we detected more than four times as many To obtain identifiers suitable for the enrichment tool genes under selection in the derived slavemaker T. DAVID 6.7 [75], we inferred orthology by applying a americanus (N = 211) than in the two younger slave- BLASTx between T. curvispinosus contigs and Drosoph- maker sister species T. pilagens (N = 38) and T. duloticus ila melanogaster protein sequences (dmel-all-transla- (N = 54; Fig. 2a). The host species T. ambiguus shares tion-r5.56.fasta) obtained from flybase (flybase.org). The one gene under positive selection with its sister species complete contig set was used as background and the ac- T. longispinosus, and two genes with T. curvispinosus, cording positively selected genes as test set. Further- whereas the latter two species do not share any posi- more, to obtain pathway information in form of KO tively selected genes (Fig. 2b). In slavemakers, one posi- (KEGG Orthology) assignments for the according gene tively selected gene (a hypothetical protein) is shared sets, we utilized KAAS [76], an automated annotation amongst all three species, T. americanus shares two server. The KEGG Mapper – Reconstruct Pathway tool genes each with the two other slavemakers, and T. pila- was used to obtain the associated pathways (http:// gens shares only one positively selected gene with its sis- www.genome.jp/kegg/tool/map_pathway.html). ter T. duloticus (Additional file 3). In both, hosts and

Fig. 1 RAxML obtained Maximum Likelihood phylogenetic relationship of the six slavemaker and host species with Acromyrmex echinatior as outgroup, and based on 5199 ortholog gene-clusters (ML bootstrap percentages depicted at nodes). Slavemaker species names are given in red, host species names in black. Arrows connect slavemaker-host species pairs Feldmeyer et al. BMC Evolutionary Biology (2017) 17:237 Page 5 of 11

Fig. 2 Venn diagrams depicting the number of shared and “private” positive selected genes (a + b) and corresponding pathways (c + d)in slavemakers and hosts. (T. dul = T. duloticus, T. ameri = T. americanus, T. pila = T. pilagens, T. curvi = T. curvispinosus, T. longi = T. longispinosus, T. ambi = T. ambiguus) slavemakers, the number of shared genes under positive in hosts and also in slavemakers, except between the selection among species is less than one would expect by sister species T. pilagens and T. duloticus (Additional chance (Additional file 4: Table S1). Moreover, we com- file 4: Table S2). The eight pathways shared among pared the positively selected genes to differentially hosts are the metabolic pathway, biosynthesis of sec- expressed genes from an accompanying gene expression ondary metabolites, biosynthesis of antibiotics, p53 analysis [64] based on the same host and slavemaker signalling, PI3K-Akt signalling, Wnt signalling, thyroid transcriptomes. Six genes in hosts and 36 in slavemakers hormone signalling and the longevity regulating path- appeared in both, the differential expression and the way. In slavemakers, four pathways were shared positive selection analyses (Table 1). None of the posi- amongst all three species including metabolic path- tively selected gene sets per species, in slavemakers or ways, biosynthesis of secondary metabolites, biosyn- hosts, were enriched for any functional category. thesis of antibiotics and endocytosis. The comparison of pathways associated with selected genes indicates that selected genes between species not only belong to different functional cat- Slavemaker-host pairs egories, but also to many different pathways (Add- Genes under selection in slavemaker-host pairs should itional file 5). In slavemakers, we identified 128 show little overlap due to their coevolution, as differ- different pathways amongst the positively selected ent traits are under selection in slavemakers and genes, the majority (77%) of which were also species- hosts. An exception could be cuticular hydrocarbon 2 specific (χ1 = 73.26, p < 0.001; Fig. 2c). The genes posi- genes, when slavemakers try to mimic host profiles tively selected in hosts belong to 102 different path- [77] and utilize the same genes as their closely related ways, the majority (70%) of which were species- hosts. A more important cause of overlap in selected 2 specific (χ1 = 29.82, p < 0.001; Fig. 2d). Nevertheless, genes in slavemaker and host pairs might be, that more pathways were shared than expected by chance both species inhabit the same habitat and therefore Feldmeyer et al. BMC Evolutionary Biology (2017) 17:237 Page 6 of 11

Table 1 List of genes which emerged as candidates in this study (signatures of positive selection), as well as an accompanying study contrasting gene expression patterns during a raid versus no raid behaviours [64] T. ambiguus T. curvispinosus T. longispinosus hypothetical protein G5I_08161 Trypsin-7 Leukotriene A-4 hydrolase hypothetical protein SINV_10379 Suppressor of tumorgenicity protein 14 Trypsin-7

T. duloticus T. pilagens T. americanus hypothetical protein SINV_06866 Putative inorganic phosphate Putative inorganic phosphate cotransporter cotransporter hypothetical protein SINV_03497 Paired amphipathic helix protein Sin3a Aminopeptidase N Alpha-catulin Zinc transporter ZIP1 RING finger protein 17 Thyrotropin-releasing hormone-degrading ectoenzyme Uncharacterized protein Sugar transporter ERD6-like 7 hypothetical protein SINV_12600 hypothetical protein SINV_09653 Matrix metalloproteinase-14 Pleckstrin-like protein domain-containing family M Kelch-like protein 10 hypothetical protein G5I_08549 member 2 F-box/LRR-repeat protein 20 Fatty acyl-CoA reductase 1 Receptor-type tyrosine-protein phosphatase beta hypothetical protein SINV_03929 hypothetical protein EAI_01741 hypothetical protein SINV_02546 Major facilitator superfamily domain-containing protein 6 Cytochrome b5 Sphingomyelin phosphodiesterase hypothetical protein SINV_09234 Zinc finger protein jing-like protein hypothetical protein G5I_14818 Elongation of very long chain fatty acids protein Putative ATP-dependent RNA helicase DDX23 Circadian clock-controlled protein Trypsin-7 hypothetical protein SINV_04023 Circadian clock-controlled protein adapt to the same environmental conditions. We in- as expected by chance in the other two parasite-host vestigated the number and functions of positively se- pairs (Additional file 4: Table S3 and S4). Moreover, lected genes and pathways between each slavemaker- we tested whether slavemakers share more genes with host pair in order to make inferences on local adapta- their preferred host in contrast to the other host spe- tion. We found between none and four shared posi- cies. T. duloticus shares more genes with T. longispi- tively selected genes (Fig. 3a-c; Additional file 6), and nosus (n = 5) in comparison to its preferred host T. 2 7–23 shared pathways between pairs (Fig. 3d-f). The curvispinosus (n =0) (X1 = 4.811, p = 0.028). In all number of shared genes and pathways was higher other cases, the number of shared genes between than expected in the pair including the most diverged preferred host and the other species did not differ parasite species T. americanus – T. longispinosus,and (results not shown).

abc

def

Fig. 3 Venn diagrams depicting shared and “private” selected genes (a-c) and pathways (d-f) between host and slavemakers species pairs. For species name abbreviations please refer to Fig. 2 Feldmeyer et al. BMC Evolutionary Biology (2017) 17:237 Page 7 of 11

Discussion homologs, and function as co-chaperons. They are in- The slavemaker lifestyle evolved several times independ- volved in stress response in humans [81], and could thus ently in ants, with a hotspot of slavery evolution in the play a role during stressful slave raids into fiercely genus Temnothorax [37]. As slavemakers and their hosts defending host colonies. are engaged in a constant co-evolutionary arms-race, the Tachykinin is positively selected on the branches lead- evolution of ant slavery is tightly linked to the evolution ing to hosts and might thus be a candidate gene for the of behaviour, physiology and morphology in their hosts host lifestyle. Tachykinin has been linked to aggressive [78–80]. The focus of our study was the genomic basis behaviour in Drosophila [82 and authors therein], and of the (co-) evolution of the slavemaker and host life- recently also in ants [83]. Temnothorax hosts need to style. We were thus interested in identifying genes with defend their nest aggressively, not only against intra- signatures of positive selection in slavemakers and hosts specific intruders, but particularly against slavemakers. respectively. Furthermore, we asked whether or not, and Thus selection might specifically act on this gene in at which organisational level, the three slavemaker/host hosts. This is further corroborated by the fact that intra- species show signs of genetic convergence; or whether specific aggression increases with the prevalence of the each species follows its own specific evolutionary slavemaker T. americanus in the population [54]. trajectory. In comparison to an accompanying gene expression analysis including the same six host and slavemaker se- Positively selected genes quence data [64], we found six genes to be both, differ- We identified twice as many genes under positive selec- entially expressed and under positive selection in hosts tion on branches leading to slavemakers as compared to and 36 in slavemakers. These genes are thus prime can- host branches. This finding is in line with our expect- didates for the evolution of the slavemaker and host life- ation that the derived slavemaker mode of life should style. Firstly, they are directly involved in raiding have led to the selection of more genes in comparison to behaviour. Secondly, they show signs of positive selec- the ancestral host lifestyle. However, based on the num- tion. Among these genes Trypsin-7 was identified in two ber of species-specific selected genes it becomes evident host species and the slavemaker T. americanus. Trypsin- that 70% of the positively selected genes in slavemakers 7 is known for its function in digestion, e.g. it is blood can be assigned to T. americanus only. All other slave- meal induced in Anopheles gambiae, and may also play a maker and host species have comparable and lower role in host seeking behaviour [84]. It may thus be in- numbers of positively selected genes. T. americanus is volved in host seeking behaviour in slavemakers and sla- the most distantly related species in this taxon and its vemaker detection in hosts. Endogenous daily (circadian) behaviour and morphology are most derived from the and annual (circannual) rhythms serve as biological other species. T. americanus workers have large square clocks that provide the major basis for timing in most heads, which make them easily distinguishable from organisms [85]. Annual timing mechanisms regulate sea- their hosts. They do not engage in normal worker be- sonal timing of reproduction, moult, and hibernation [86 haviour, such as brood care or foraging, and are so and authors therein]. Positive selection on the circadian dependent upon enslaved hosts that T. americanus will clock controlled protein, in slavemakers compared to starve to death if not fed. During raids, they manipulate hosts, suggests that this gene may regulate the aberrant host behaviour via the release of glandular secretions activity patterns of slavemaker workers. Slavemakers are [11, 41, 46], but never use their stinger, which is a typical only active during raiding season in summer, and are behaviour for other Temnothorax hosts during aggres- taken care of for the rest of the year by the slaves [39, sive interactions [7, 41, 48]. Foraging and brood care are 87]. This changed activity pattern might thus be mani- standard behavioural repertoires in the hosts, and in the fested by changes in the circadian rhythm. Two more lack of slaves, will still be performed by T. duloticus [41, 44] genes of interest with possible direct link to the slave- and T. pilagens (pers. observation) slavemakers to some ex- maker evolution are “ Elongation of very long chain fatty tent. In addition to many lifestyle differences, the longer acids protein” in T. duloticus and Fatty acyl-CoA reduc- evolutionary history with the possibility for co- and counter tase 1 in T. pilagens. Both genes could be involved in adaptations to T. longispinosus and its host ancestor, might the synthesis of cuticular hydrocarbons and thus might explain the large number of positively selected genes in T. play a role in the avoidance of host recognition [88]. In- americanus in contrast to the other species. deed, a recent study on the cuticular hydrocarbon pro- Within the 309 positively selected genes on the files of the same six species, revealed that slavemakers branches to slavemakers, we were able to identify several show consistently different chemical profiles than the candidate genes with a possible link to their slavemaker three host species [89]. A recent switch to a parasitic lifestyle. Amongst these, three different DNAJ-like pro- lifestyle could thus have led to selection on genes under- tein subfamily members, which are heat shock protein lying hydrocarbon synthesis. Feldmeyer et al. BMC Evolutionary Biology (2017) 17:237 Page 8 of 11

In slavemakers only four pathways are shared among occur in the same environment, in comparison to the species. Three of these (metabolic pathways, biosynthesis other four species which are from different locales. of secondary metabolites, and biosynthesis of antibiotics) were also identified in the hosts, and only endocytosis is Conclusions slavemaker specific. In hosts, the genes with signatures Our positively selected gene analyses revealed several of positive selection belong to 102 different pathways, candidate genes with a possible link to the slavemaker eight of which are shared among the three species. The PI3K-Akt pathway is part of the mTOR pathway regulat- lifestyle, which are involved in the cuticular hydrocarbon profile composition, thus species recognition, and the ing the cell cycle, and also known for its function in lon- aberrant activity pattern of slavemaker workers. To ver- gevity [90]. Furthermore the longevity regulating pathway is shared among all three host species, reinfor- ify the functional and phenotypic importance of these candidates will now be the next step. cing the importance of longevity within these host spe- Furthermore, the results show little overlap of selected cies. Despite their small body and colony size Temnothorax ants are quite long-lived with workers liv- genes between species. On the pathway level however, we find higher congruence between species than ex- ing up to a few years and queens over two decades [91]. pected, even though the majority of selected pathways Social Hymenopterans are known for a change in the longevity-fecundity trade-off with queens being both remain species specific. Furthermore, the genes under positive selection belong to a wide variety of functions, long-lived and highly fecund compared to the short- as indicated by negative results in the enrichment ana- lived sterile workers [92–97]. We identified two pathways, which may play a role in morphological differences lyses. The same pattern was identified in social parasitic cape honeybees [37]. It thus seems that the evolution of between the species and their adaptive divergence. The social parasites, including slavemakers is a broad encom- Wnt signalling pathway is known for its role in regulating key events during embryonic patterning and mor- passing process with species-specific evolutionary trajectories, based on selection in many genes with different phogenesis [98], and the thyroid hormone signalling functionality and pathway affiliation. Our results support pathway (in humans) is involved in the regulation of growth development and metabolism. The latter has the hypothesis that evolution is the unrepeatable result of stochastic events with highly contingent effects [101]. been shown to play a role in the adaptive divergence of sticklebacks [99]. Additional files

Slavemaker – Host pairs Additional file 1: Information on sample collection sites and year for Besides determining similarities in possible selection each species. (DOCX 112 kb) pressures within slavemakers and within hosts, we add- Additional file 2: Summary of read counts and contig information per itionally investigated similarities in slavemaker-host species. (DOCX 14 kb) Additional file 3: Codeml results on selected gene identified on pairs, because of their shared environment. We hypothe- branches leading to slavemakers and hosts, as well as each single sized that genes shared by both slavemakers and hosts species. Plus information on shared and private selected genes. (XLSX with signatures of positive selection might give indica- 135 kb) tion on local environmental selection pressures; though Additional file 4: Results of randomisation statistics. (DOCX 16 kb) they could also represent genes involved in the co- Additional file 5: KEGG pathways assigned to positively selected genes evolutionary arms race. However, on the gene level there per branch. (XLSX 21 kb) Additional file 6: Summary shared selected genes between slavemaker- is hardly any congruence between slavemaker and host host species pairs. (XLSX 39 kb) pairs (0–4 overlapping genes). On the pathway level, some of the above mentioned candidates appear, such as thyroid hormone signalling pathway in T. ambiguus and Abbreviations T. pilagens, PI3K-Akt and Wnt signalling pathway be- L:D: Light: dark; dN/dS: Ratio of non-synonymous to synonymous mutations; mtDNA: Mitochondrial DNA tween T. americanus and T. longispinosus. In the T. duloticus – T. curvispinosus pair we identified circadian Acknowledgements rhythm as well as the FoxO signalling pathway. Among The authors would like to thank several anonymous reviewers for helpful others the latter coordinates the response to environ- comments on the content of this study. mental changes, including metabolic stress (starvation) and oxidative stress [100]. Hence, these two pathways Funding may give evidence for environmental selection pressures, This project was funded by a grant to Barbara Feldmeyer and Susanne Foitzik (DFG FE 1333/3–1 and FO 298/17–1) and support from the E. N. e.g. temperature, seasonality, or food availability, experi- Huyck Preserve, Rensselearville NY. Collection permits were obtained at the enced by T. duloticus and T. curvispinosus which co- Parks and Preserves. Feldmeyer et al. BMC Evolutionary Biology (2017) 17:237 Page 9 of 11

Availability of data and materials 15. Tellier A, Brown JKM. Stability of genetic polymorphism in host–parasite All data analysed during this study are included in various Additional files interactions. Proc R Soc B Biol Sci. 2007;274:809–17. (see below) and the contig ortholog clusters are available from NCBIO 16. González-Tortuero E, Rusek J, Turko P, Petrusek A, Maayan I, Piálek L, BioProject GSE95604 (GSE95604_2017July_TemnoClusters.fas-gb.tar.gz). Tellenbach C, Gießler S, Spaak P, Wolinska J. Daphnia parasite dynamics across multiple Caullerya epidemics indicate selection against common Authors’ contributions parasite genotypes. Zoology. 2016;119:314–21. DE and AA assembled the transcriptomes and created the ortholog clusters. 17. Papkou A, Gokhale CS, Traulsen A, Schulenburg H. Host – parasite DE and BF conducted the selected genes analyses. BF drafted the coevolution: why changing population size matters. Zoology. 2016;119: manuscript, all authors read and edited the manuscript and approved it 330–8. before submission. 18. Gandon S, Michalakis Y. Local adaptation, evolutionary potential and host- parasite coevolution: interactions between migration, mutation, population – Ethics approval and consent to participate size and generation time. J Evol Biol. 2002;15:451 62. Not applicable. 19. Wolinska J, King KC. Environment can alter selection in host–parasite interactions. Trends Parasitol. 2009;25:236–44. Consent for publication 20. Bose J, Kloesener MH, Schulte RD. Multiple-genotype infections and their – Not applicable. complex effect on virulence. Zoology. 2016;119:339 49. 21. Kurze C, Routtu J, Moritz RFA. Parasite resistance and tolerance in honeybees at the individual and social level. Zoology. 2016;119:290–7. Competing interests 22. Joop G, Vilcinskas A. Coevolution of parasitic fungi and insect hosts. The authors declare that they have no competing interests. Zoology. 2016;119:350–8. 23. Croze M, Živković D, Stephan W, Hutter S. Balancing selection on immunity Publisher’sNote genes: review of the current literature and new analysis in Drosophila – Springer Nature remains neutral with regard to jurisdictional claims in melanogaster. Zoology. 2016;119:322 9. published maps and institutional affiliations. 24. Brooke M, Davies NB. Egg mimicry by cuckoos Cuculus canorus in relation to discrimination by hosts. Nature. 1988;335:630–2. Author details 25. Davies NB. Cuckoos, cowbirds and other cheats. London: T. & A. D. Poyser; 1Senckenberg Biodiversity and Climate Research Centre (BiK-F), Molecular 2000. Ecology, Senckenberganlage 25, 60325 Frankfurt am Main, Germany. 26. Yamauchi A. Theory of evolution of nest parasitism in birds. Am Nat. 1995; 2Evolutionary Biology and Ecology, University of Freiburg, Hauptstrasse 1, 145:434–56. 79104 Freiburg, Germany. 3Institute of Organismic and Molecular Evolution, 27. Leppänen J, Seppä P, Vepsäläinen K, Savolainen R. Genetic divergence Johannes Gutenberg University Mainz, Johannes von Müller Weg 6, 55128 between the sympatric queen morphs of the ant Myrmica rubra.Mol Mainz, Germany. Ecol. 2015;24:2463–76. 28. Hamilton WJ, Orians GH. Evolution of brood parasitism in altricial birds. Received: 19 July 2017 Accepted: 16 November 2017 Condor. 1965;67:361–82. 29. Fanelli D, Henshaw M, Cervo R, Turillazzi S, Queller DC, Strassmann JE. The social parasite wasp Polistes atrimandibularis does not form host – References races. J Evol. 2005;18:1362 7. 1. Poulin R, Morand S. The diversity of parasites. Q Rev Biol. 2000;75:277–93. 30. Darwin C. On the origin of species. Cambridge, Mass: Harvard University 2. Viney M, Cable J. Macroparasite life histories. Curr Biol. 2011;21:R767–74. Press; 1859. 3. Buschinger A: Bettel, Raub und Mord: aus dem Leben sozialparasitischer 31. Zhang L, Zhou Z, Guo Q, Fokkens L, Miskei M, Pócsi I, Zhang W, Chen M, Ameisen. Verh Westdtsch Entomol 1993, 1993:7–20. Wang L, Sun Y, Donzelli BGG, Gibson DM, Nelson DR, Luo J-G, Rep M, Liu H, 4. Davies NB, Bourke AFG, Brooke M de L. Cuckoos and parasitic ants: Yang S, Wang J, Krasnoff SB, Xu Y, Molnár I, Lin M. Insights into adaptations Interspecific brood parasitism as an evolutionary arms race. Trends Ecol to a near-obligate nematode endoparasitic lifestyle from the finished Evol. 1989;4:274–8. genome of Drechmeria coniospora. Sci Rep. 2016;6:23122. 5. de Meeûs T, Michalakis Y, Renaud F. Santa Rosalia revisited: or why are there 32. Skippington E, Barkman TJ, Rice DW, Palmer JD. Miniaturized mitogenome so many kinds of parasites in “the garden of earthly delights”? Parasitol of the parasitic plant Viscum scurruloideum is extremely divergent and Today. 1998;14:10–3. dynamic and has lost all nad genes. PNAS. 2015:112:E3515–24. 6. Dawkins R, Krebs JR. Arms races between and within species. Proc R Soc 33. Jackson AP. Preface: The evolution of parasite genomes and the origins of Lond B Biol Sci. 1979;205:489–511. parasitism. Parasitology. 2015;142:S1–5. 7. Foitzik S, de Heer CJ, Hunjan DN, Herbers JM. Coevolution in host-parasite 34. Poulin R, Randhawa HS. Evolution of parasitism along convergent lines: systems: behavioural strategies of slave-making ants and their hosts. Proc R from ecology to genomics. Parasitology. 2015;142:S6–S15. Soc Lond B. 2001;268:1139–46. 35. Smith CR, Helms Cahan S, Kemena C, Brady G, Yang W, Bornberg E, Smith 8. Calcagno V, Dubosclard M, de Mazancourt C. Rapid exploiter-victim CR, Cahan SH, Kemena C, Eriksson T, Gadau J, Helmkampf M, Gotzek D, coevolution: the race is not always to the swift. Am Nat. 2010;176:198–211. Okamoto M, Suarez AV, Mikheyev A. How do genomes create novel 9. Kilner RM, Langmore NE. Cuckoos versus hosts in insects and birds: phenotypes ? Insights from the loss of the worker caste in ant social adaptations, counter-adaptations and outcomes. Biol Rev Camb Philos Soc. parasites. Mol Biol Evol. 2015;32:2919–31. 2011;86:836–52. 36. Goudie F, Allsopp MH, Solignac M, Beekman M, Oldroyd BP. The frequency 10. Grasso DA, Mori A, Le Moli F. Analysis of the aggression between slave and of arrhenotoky in the normally thelytokous Apis mellifera capensis worker slave-making (facultative and obligatory) ant species (Hymenoptera and the clone reproductive parasite. Insect Soc. 2015;62:325–33. Formicidae). Ethology Ecol Evol. 1992;4:81–4. 37. Wallberg A, Pirk CW, Allsopp MH, Webster MT. Identification of multiple loci 11. Jongepier E, Foitzik S: Ant recognition cue diversity is higher in the associated with social parasitism in honeybees. PLoS Genet. 2016;12:1–30. presence of slavemaker ants. Behav Ecol. 2015;27:304–11. 38. Beibl J, Stuart RJ, Heinze J, Foitzik S. Six origins of slavery in formicoxenine 12. Ortolani I, Zechini L, Turillazzi S, Cervo R. Recognition of a paper wasp social ants. Insect Soc. 2005;52:291–7. parasite by its host: evidence for a visual signal reducing host 39. Herbers JM, Foitzik S, Collins F. The ecology of slavemaking ants and their aggressiveness. Anim Behav. 2010;80:683–8. hosts in north temperate forests. Ecology. 2002;83:148–63. 13. Barrett LG, Thrall PH, Burdon JJ, Linde CC. Life history determines genetic 40. Seifert B, Kleeberg I, Feldmeyer B, Pamminger T, Jongepier E, Foitzik S. Temnothorax structure and evolutionary potential of host–parasite interactions. Trends pilagens sp. n. – a new slave-making species of the tribe Formicoxenini from North Ecol Evol. 2008;23:678–85. America (Hymenoptera, Formicidae). Zookeys. 2014;368:65–77. 14. Strauß JF, Crain P, Schulenburg H, Telschow A. Experimental evolution in 41. Alloway TM. Raiding behaviour of two species of slave-making ants, silico: a custom-designed mathematical model for virulence evolution of Harpagoxenus americanus (Emery) and Leptothorax duloticus Wesson Bacillus thuringiensis. Zool (Jena). 2016;119:359–65. (Hymenoptera: Formicidae). Anim Behav. 1979;27:202–10. Feldmeyer et al. BMC Evolutionary Biology (2017) 17:237 Page 10 of 11

42. Foitzik S, Achenbach A, Brandt M. Locally adapted social parasite affects 69. Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ. Basic local alignment density, social structure, and life history of its ant hosts. Ecology. 2009;90: search tool. J Mol Biol. 1990;215:403–10. 1195–206. 70. Katoh K, Standley DM. MAFFT multiple sequence alignment software 43. Johnson CA, Herbers JM. Impact of parasite sympatry on the geographic version 7: improvements in performance and usability. Mol Biol Evol. 2013; mosaic of coevolution. Ecology. 2006;87:382–94. 30:772–80. 44. Wilson EO. Leptothorax duloticus and the beginnings of slavery in ants. 71. Abascal F, Zardoya R, Posada D. ProtTest: selection of best-fit models of Evolution. 1975;29:108–19. protein evolution. Bioinformatics. 2005;21:2104–5. 45. Stuart RJ, Alloway TM. Behavioural evolution and domestic degeneration in 72. Stamatakis A. RAxML-VI-HPC: maximum likelihood-based phylogenetic obligatory slave-making ants (Hymenoptera: Formicidae: Leptothoracini). analyses with thousands of taxa and mixed models. Bioinformatics. 2006;22: Anim Behav. 1985;33:1080–8. 2688–90. 46. Brandt M, Heinze J, Schmitt T, Foitzik S. Convergent evolution of the 73. Yang Z. PAML 4: Phylogenetic analysis by maximum likelihood. Mol Biol Dufour’s gland secretion as a propaganda substance in the slave-making Evol. 2007;24:1586–91. ant genera Protomognathus and Harpagoxenus. Insect Soc. 2006;53:291–9. 74. Benjamni Y, Hochberg Y, Benjamini Y. Controlling the false discovery rate: a 47. Jongepier E, Kleeberg I, Foitzik S. The ecological success of a social practical and powerful approach to multiple testing. J R Stat Soc Ser B- parasite increases with manipulation of collective host behaviour. J Evol Methodological. 1995;57:289–300. Biol. 2015;12:2152–62. 75. Huang DW, Sherman BT, Lempicki RA. Bioinformatics enrichment tools: 48. Kleeberg I, Foitzik S. The placid slavemaker: avoiding detection and paths toward the comprehensive functional analysis of large gene lists. conflict as an alternative, peaceful raiding strategy. Behav Ecol Nucleic Acids Res. 2009:1–13. Sociobiol. 2016;70:27–39. 76. Moriya Y, Itoh M, Okuda S, Yoshizawa AC, Kanehisa M. KAAS: an automatic 49. Hare JF, Alloway TM. Prudent Protomognathus and despotic Leptothorax genome annotation and pathway reconstruction server. Nucleic Acids Res. duloticus: differential costs of ant slavery. PNAS. 2001;98:12093–6. 2007;35:W182–5. 50. Brandt M, Foitzik S. Community context and specialization influence 77. Achenbach A, Witte V, Foitzik S. Brood exchange experiments and chemical coevolution between a slavemaking ant and its hosts. Ecology. 2004;85: analyses shed light on slave rebellion in ants. Behav Ecol. 2010;21:948–56. 2997–3009. 78. Gross P. Insect behavioral and morphological defenses against parasitoids. 51. Pamminger T, Scharf I, Pennings PS, Foitzik S. Increased host aggression as Annu Rev Entomol. 1993;38:251–73. an induced defense against slave-making ants. Behav Ecol. 2011;22:255–60. 79. Hart BL. Behavioural defence. In: Clayton DH, Moore J, editors. Host-parasite 52. Scharf I, Pamminger T, Foitzik S. Differential response of ant colonies to evolution. General principles and avian models. Oxford: Oxford University intruders: attack strategies correlate with potential threat. Ethology. 2011; Press; 1997. p. 59–77. 117:731–9. 80. Sorci G. Immunity, resistance and tolerance in bird – parasite interactions. 53. Kleeberg I, Pamminger T, Jongepier E, Papenhagen M, Foitzik S. Forewarned Parasite Immunol. 2013;35:350–61. is forearmed: aggression and information use determine fitness costs of 81. Terada K, Mori M. Human DnaJ homologs dj2 and dj3, and bag-1 are slave raids. Behav Ecol. 2014;25:1058–63. positive cochaperones of hsc70. J Biol Chem. 2000;275:24728–34. 54. Kleeberg I, Jongepier E, Job S, Foitzik S. Geographic variation in social 82. Pavlou HJ, Neville MC, Goodwin SF. Aggression: Tachykinin is all the rage. parasite pressure predicts intraspecific but not interspecific aggressive Curr Biol. 2014;24:R243–4. responses in hosts of a slavemaking ant. Ethology. 2015;121:1–9. 83. Howe J, Schiøtt M, Boomsma JJ. Tachykinin expression levels correlate with 55. Jongepier E, Kleeberg I, Job S, Foitzik S. Collective defence portfolios of ant caste-specific aggression in workers of the leaf-cutting ant Acromyrmex hosts shift with social parasite pressure. Proc R Soc B. 2014;281:20140225. echinatior. Front Ecol Evol. 2016;4(May):1–12. 56. Conte GL, Arnegard ME, Peichel CL, Schluter D. The probability of genetic 84. Muller HM, Catteruccia F, Vizioli J, Dellatorre A, Crisanti A. Constitutive and parallelism and convergence in natural populations. Proc R Soc B. 2012;279: blood meal-induced trypsin genes in Anopheles gambiae. Exp Parasitol. 5039–47. 1995;81:371–85. 57. Martin A, Orgogozo V. The loci of repeated evolution: a catalog of genetic 85. Gwinner E. Circadian and circannual programmes in avian migration. J Exp hotspots of phenotypic variation. Evolution. 2013;67:1235–50. Biol. 1996;199:39–48. 58. Elmer KR, Meyer A. Adaptation in the age of ecological genomics: insights 86. Wikelski M, Martin LB, Scheuerlein A, Robinson MT, Robinson ND, Helm B, from parallelism and convergence. Trends Ecol Evol. 2011;26:298–306. Hau M, Gwinner E. Avian circannual clocks: adaptive significance and 59. Stern DL. The genetic causes of convergent evolution. Nat Rev Genet. 2013; possible involvement of energy turnover in their proximate control. Philos 14:751–64. Trans R Soc Lond B. 2008;363:411–23. 60. Soria-Carrasco V, Gompert Z, Comeault AA, Farkas TE, Parchman TL, 87. Pohl S, Foitzik S. Slave-making ants prefer larger, better defended host Johnston JS, Buerkle CA, Feder JL, Bast J, Schwander T, Egan SP, Crespi BJ, colonies. Anim Behav. 2011;81:61–8. Nosil P. Stick insect genomes reveal natural selection’s role in parallel 88. Blomquist GJ, Bagnères A-G. Insect hydrocarbons biology, biochemistry and speciation. Science. 2014;344:738–42. ecology. Cambridge, UK: Cambridge University Press; 2010. 61. Lobkovsky AE, Koonin EV. Replaying the tape of life: quantification of the 89. Kleeberg I, Menzel F, Foitzik S. The influence of slavemaking lifestyle, caste predictability of evolution. Front Genet. 2012;3:246. and sex on chemical profiles in Temnothorax ants: insights into the 62. Bolger AM, Lohse M, Usadel B. Trimmomatic: a flexible trimmer for Illumina evolution of cuticular hydrocarbons. Proc R Soc B. 2017;284:20162249. sequence data. Bioinformatics. 2014;30:2114–20. 90. Hay N. Interplay between FOXO, TOR, and Akt. Biochim Biophys Acta. 2011; 63. Chevreux B, Wetter T, Suhai S. Genome sequence assembly using trace 1813:1965–70. signals and additional sequence information. Comput Sci Biol Proc Ger Conf 91. Keller L, Genoud M. Extraordinary lifespans in ants: a test of evolutionary Bioinforma. 1999;99:45–56. theories of ageing. Nature. 1997;389:958–60. 64. Alleman A, Feldmeyer B, Foitzik S. Comparative analyses of co-evolving 92. Heinze J, Frohschammer S, Bernadou A. Queen life-span and total host-parasite associations reveal unique expression patterns and pathways reproductive success are positively associated in the ant Cardiocondyla cf. underlying slavemaker raiding and host defensive behavior in ants. in prep. kagutsuchi. Behav Ecol Sociobiol. 2013;67:1555–62. 65. Min XJ, Butler G, Storms R, Tsang A. OrfPredictor: predicting protein-coding 93. Oettler J, Schrempf A. Fitness and aging in Cardiocondyla obscurior ant regions in EST-derived sequences. Nucleic Acids Res. 2005;33:W677–80. queens. Curr Opin Insect Sci. 2016;16:58–63. 66. Li L, Stoeckert CJ, Roos CJ. OrthoMCL: identification of ortholog groups for 94. Rodrigues MA, Flatt T. Endocrine uncoupling of the trade-off between eukaryotic genomes. Genome Res. 2003;13:2178–89. reproduction and somatic maintenance in eusocial insects. Curr Opin Insect 67. Castresana J. Selection of conserved blocks from multiple alignments for Sci. 2016;16:1–8. their use in phylogenetic analysis. Mol Biol Evol. 2000;17:540–52. 95. Rueppell O, Aumer D, Moritz RFA. Ties between ageing plasticity and 68. Nygaard S, Zhang G, Schiøtt M, Nygaard S, Zhang G, Schiøtt M, Li C, Wurm reproductive physiology in honey bees (Apis mellifera) reveal a positive Y, Hu H, Zhou J, Ji L, Qiu F, Rasmussen M, Pan H, Hauser F, Krogh A, relation between fecundity and longevity as consequence of advanced Grimmelikhuijzen CJP, Wang J. The genome of the leaf-cutting ant social evolution. Curr Opin Insect Sci. 2016;16:64–8. Acromyrmex echinatior suggests key adaptations to advanced social life and 96. Toth AL, Sumner S, Jeanne RL. Patterns of longevity across a sociality fungus farming. Genome Res. 2011;21:1339–48. gradient in vespid wasps. Curr Opin Insect Sci. 2016;16:28–35. Feldmeyer et al. BMC Evolutionary Biology (2017) 17:237 Page 11 of 11

97. Negroni MA, Jongepier E, Feldmeyer B, Kramer BH, Foitzik S. Life history evolution in social insects: a female perspective. Curr Opin Insect Sci. 2016; 16:51–7. 98. Komiya Y, Habas R. Wnt signal transduction pathways. Organ. 2008;4:68–75. 99. Kitano J, Lema SC, Luckenbach JA, Mori S, Kawagishi Y, Kusakabe M, Swanson P, Peichel CL. Report adaptive divergence in the thyroid hormone signaling pathway in the stickleback radiation. Curr Biol. 2010;20:2124–30. 100. Eijkelenboom A, Burgering BMT. FOXOs: signalling integrators for homeostasis maintenance. Nat Rev Mol Cell Biol. 2013;14:83–97. 101. Gould SJ. Wonderful life: the burgess shale and the nature of history. New York: W.W. Norton & Company; 1989.

Submit your next manuscript to BioMed Central and we will help you at every step:

• We accept pre-submission inquiries • Our selector tool helps you to ﬁnd the most relevant journal • We provide round the clock customer support • Convenient online submission • Thorough peer review • Inclusion in PubMed and all major indexing services • Maximum visibility for your research

Submit your manuscript at www.biomedcentral.com/submit